DI pipe thrust - look at the web site for DIPRA (ductile iron pipe research assoc) of the USA. They provide a program for calculating the length required (for thrust restraint) so that thrust blocks are not required. The theory was developed in about 1976 & has been in use since that date by DIPRA.
The pipes are 'harnessed' by using special joints for the required length. In Australia, a TytonLok joint can be used (Tyco DI pipes) & in the US there are several manufacturers of restrained joints. The thrust force is taken the same way as for piles (but in a horizontal direction), by using a combination of both skin friction (around the pipe) & end bearing (of the pipe & fitting).
Be careful about using AWWA M11 - steel pipe design & installation (& the Tyco manual in Australia), because they use an anchorage length virtually based on the vertical mass times the coefficient of friction. This is not representative of 'true' anchorage because it only takes into account 'shear box' type of friction. This is not how a pipe anchorage resistance is developed. The result will probably be several times the length calculated by the DIPRA method. The DIPRA method has stood the test of time.
To Ussuri
In the water industry in my country, flanged joints are used VERY rarely on pipelines except for valves (& even then they are installed with a rubber ring dismantling joint). Even bends are RRJ. The water industry does not like many items that can corrode such as flanges, bolts & nuts. A flanged joint (or pipe) is more expensive, takes longer to put together, & is a source of corrosion.
Flanged joints are used on bends for above ground situations (pump stations & bore heads etc) because they allow not only easy dismantling, but can be replaced without disrupting the pipeline.
BarryEng